EXPLAINABLE MACHINE LEARNING¶

@Author Gabriel Schurr, Ilyesse Hettenbach

IMPORTS¶

In [ ]:
import os
import random

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
import torch
import torch.nn as nn
import torch.optim as optim
from dotenv import load_dotenv
from PIL import Image
from torch.utils.data import DataLoader, Dataset
from torchsummary import summary
from torchvision import models, transforms
from torchvision.datasets import ImageFolder
from tqdm.notebook import tqdm

from torch_utils import get_loaders, transform

load_dotenv()
Out[ ]:
True
In [ ]:
DATAPATH = str(os.getenv("DATAPATH"))
P_LABELS = os.path.join(DATAPATH, "images_labels.txt")
DATAPATH = os.path.join(DATAPATH, "animals")
print(f'Path to images: {DATAPATH}')
print(f'Path to labels: {P_LABELS}')

# P_LABELS = "D:\\Database\\animals\\original\\images_labels.txt"
# DATAPATH = "D:\\Database\\animals\\original\\animals"

DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
# DEVICE = 'cpu'

BATCH_SIZE = 256
EPOCHS = 3
LR = 0.0007
RANDOM_SEED = 42
NUM_WORKERS = 0

print(f"Device: {DEVICE}")

Path to images: W:/Workspaces/Python/Python-Playground/src/Studium/ExplainableMLTrain/src/animal-image-dataset-90-different-animals/animals\animals
Path to labels: W:/Workspaces/Python/Python-Playground/src/Studium/ExplainableMLTrain/src/animal-image-dataset-90-different-animals/animals\images_labels.txt
Device: cuda
In [ ]:
do_train = False
do_checkpoint = False
do_load_model = True
model_path = "best_model.pth"

EDA¶

In [ ]:
data = []
with open(P_LABELS, 'r') as f:
    for line in f:
        image_path, label = line.strip().split(' ')
        data.append({'image_path': image_path, 'label': label})
animal_df = pd.DataFrame(data)
animal_df.head()
Out[ ]:
image_path label
0 W:/Workspaces/Python/Python-Playground/src/Stu... antelope
1 W:/Workspaces/Python/Python-Playground/src/Stu... antelope
2 W:/Workspaces/Python/Python-Playground/src/Stu... antelope
3 W:/Workspaces/Python/Python-Playground/src/Stu... antelope
4 W:/Workspaces/Python/Python-Playground/src/Stu... antelope
In [ ]:
animal_df.describe()
Out[ ]:
image_path label
count 5375 5375
unique 5375 90
top W:/Workspaces/Python/Python-Playground/src/Stu... antelope
freq 1 60
In [ ]:
fig, axes = plt.subplots(2, 3, figsize=(10, 6))
for i, ax in enumerate(axes.flat):
    random_index = random.randint(0, len(animal_df)-1)
    img = Image.open(animal_df['image_path'][random_index])
    label = animal_df['label'][random_index]
    ax.imshow(img)
    ax.set_title(label)
    ax.axis('off')
plt.tight_layout()
plt.show()
No description has been provided for this image
In [ ]:
# Visualize class distribution
plt.figure(figsize=(20, 6))
sns.countplot(x='label', data=animal_df)
plt.title('Class Distribution')
plt.xticks(rotation=90)
plt.show()
No description has been provided for this image

MODEL¶

In [ ]:
class CustomResNet18(nn.Module):
    def __init__(self, num_classes=90):
        super(CustomResNet18, self).__init__()
        self.output = None
        self.resnet = models.resnet18(weights='IMAGENET1K_V1') # 'IMAGENET1K_V1'
        # for param in self.resnet.parameters():
        #     param.requires_grad = False
        num_features = self.resnet.fc.in_features
        self.resnet.fc = nn.Sequential(
            nn.Linear(num_features, num_classes)
        )

    def forward(self, x):
        self.output = self.resnet(x)
        return self.output

    def freeze_backbone(self):
        for param in self.resnet.parameters():
            param.requires_grad = False

        for param in self.resnet.fc.parameters():
            param.requires_grad = True

    def unfreeze_backbone(self):
        for param in self.resnet.parameters():
            param.requires_grad = True

model = CustomResNet18()
if do_load_model and os.path.exists(model_path):
    model = torch.load(model_path)
    print("Model loaded")

summary(model, (3, 224, 224), device="cpu")

Model loaded
----------------------------------------------------------------
        Layer (type)               Output Shape         Param #
================================================================
            Conv2d-1         [-1, 64, 112, 112]           9,408
       BatchNorm2d-2         [-1, 64, 112, 112]             128
              ReLU-3         [-1, 64, 112, 112]               0
         MaxPool2d-4           [-1, 64, 56, 56]               0
            Conv2d-5           [-1, 64, 56, 56]          36,864
       BatchNorm2d-6           [-1, 64, 56, 56]             128
              ReLU-7           [-1, 64, 56, 56]               0
            Conv2d-8           [-1, 64, 56, 56]          36,864
       BatchNorm2d-9           [-1, 64, 56, 56]             128
             ReLU-10           [-1, 64, 56, 56]               0
       BasicBlock-11           [-1, 64, 56, 56]               0
           Conv2d-12           [-1, 64, 56, 56]          36,864
      BatchNorm2d-13           [-1, 64, 56, 56]             128
             ReLU-14           [-1, 64, 56, 56]               0
           Conv2d-15           [-1, 64, 56, 56]          36,864
      BatchNorm2d-16           [-1, 64, 56, 56]             128
             ReLU-17           [-1, 64, 56, 56]               0
       BasicBlock-18           [-1, 64, 56, 56]               0
           Conv2d-19          [-1, 128, 28, 28]          73,728
      BatchNorm2d-20          [-1, 128, 28, 28]             256
             ReLU-21          [-1, 128, 28, 28]               0
           Conv2d-22          [-1, 128, 28, 28]         147,456
      BatchNorm2d-23          [-1, 128, 28, 28]             256
           Conv2d-24          [-1, 128, 28, 28]           8,192
      BatchNorm2d-25          [-1, 128, 28, 28]             256
             ReLU-26          [-1, 128, 28, 28]               0
       BasicBlock-27          [-1, 128, 28, 28]               0
           Conv2d-28          [-1, 128, 28, 28]         147,456
      BatchNorm2d-29          [-1, 128, 28, 28]             256
             ReLU-30          [-1, 128, 28, 28]               0
           Conv2d-31          [-1, 128, 28, 28]         147,456
      BatchNorm2d-32          [-1, 128, 28, 28]             256
             ReLU-33          [-1, 128, 28, 28]               0
       BasicBlock-34          [-1, 128, 28, 28]               0
           Conv2d-35          [-1, 256, 14, 14]         294,912
      BatchNorm2d-36          [-1, 256, 14, 14]             512
             ReLU-37          [-1, 256, 14, 14]               0
           Conv2d-38          [-1, 256, 14, 14]         589,824
      BatchNorm2d-39          [-1, 256, 14, 14]             512
           Conv2d-40          [-1, 256, 14, 14]          32,768
      BatchNorm2d-41          [-1, 256, 14, 14]             512
             ReLU-42          [-1, 256, 14, 14]               0
       BasicBlock-43          [-1, 256, 14, 14]               0
           Conv2d-44          [-1, 256, 14, 14]         589,824
      BatchNorm2d-45          [-1, 256, 14, 14]             512
             ReLU-46          [-1, 256, 14, 14]               0
           Conv2d-47          [-1, 256, 14, 14]         589,824
      BatchNorm2d-48          [-1, 256, 14, 14]             512
             ReLU-49          [-1, 256, 14, 14]               0
       BasicBlock-50          [-1, 256, 14, 14]               0
           Conv2d-51            [-1, 512, 7, 7]       1,179,648
      BatchNorm2d-52            [-1, 512, 7, 7]           1,024
             ReLU-53            [-1, 512, 7, 7]               0
           Conv2d-54            [-1, 512, 7, 7]       2,359,296
      BatchNorm2d-55            [-1, 512, 7, 7]           1,024
           Conv2d-56            [-1, 512, 7, 7]         131,072
      BatchNorm2d-57            [-1, 512, 7, 7]           1,024
             ReLU-58            [-1, 512, 7, 7]               0
       BasicBlock-59            [-1, 512, 7, 7]               0
           Conv2d-60            [-1, 512, 7, 7]       2,359,296
      BatchNorm2d-61            [-1, 512, 7, 7]           1,024
             ReLU-62            [-1, 512, 7, 7]               0
           Conv2d-63            [-1, 512, 7, 7]       2,359,296
      BatchNorm2d-64            [-1, 512, 7, 7]           1,024
             ReLU-65            [-1, 512, 7, 7]               0
       BasicBlock-66            [-1, 512, 7, 7]               0
AdaptiveAvgPool2d-67            [-1, 512, 1, 1]               0
           Linear-68                   [-1, 90]          46,170
           ResNet-69                   [-1, 90]               0
================================================================
Total params: 11,222,682
Trainable params: 11,222,682
Non-trainable params: 0
----------------------------------------------------------------
Input size (MB): 0.57
Forward/backward pass size (MB): 62.79
Params size (MB): 42.81
Estimated Total Size (MB): 106.17
----------------------------------------------------------------
In [ ]:
animal_df["class"] = animal_df["label"].astype("category").cat.codes
class_to_animal = animal_df[['class', 'label']].drop_duplicates().set_index('class').to_dict()["label"]
animal_to_class = {v: k for k, v in class_to_animal.items()}
animal_df.head()
Out[ ]:
image_path label class
0 W:/Workspaces/Python/Python-Playground/src/Stu... antelope 0
1 W:/Workspaces/Python/Python-Playground/src/Stu... antelope 0
2 W:/Workspaces/Python/Python-Playground/src/Stu... antelope 0
3 W:/Workspaces/Python/Python-Playground/src/Stu... antelope 0
4 W:/Workspaces/Python/Python-Playground/src/Stu... antelope 0
In [ ]:
train_df = animal_df.sample(frac=0.8, random_state=42)
test_df = animal_df.drop(train_df.index)
train_df = train_df.reset_index(drop=True)
test_df = test_df.reset_index(drop=True)
In [ ]:
fig, axes = plt.subplots(1, 2, figsize=(15, 4))
sns.countplot(x='label', data=train_df, ax=axes[0])
axes[0].set_xticks([])
axes[0].set_title('Train Class Distribution')
sns.countplot(x='label', data=test_df, ax=axes[1])
axes[1].set_xticks([])
axes[1].set_title('Test Class Distribution')
plt.tight_layout()
plt.show()
No description has been provided for this image
In [ ]:
train_loader, val_loader = get_loaders(train_df, test_df, batch_size=BATCH_SIZE, num_workers=NUM_WORKERS)
In [ ]:
criterion = nn.CrossEntropyLoss()
optimizer = optim.AdamW(model.parameters(), lr=LR)

TRAINING¶

In [ ]:
if do_train:
    model.train().to(DEVICE)
    train_losses = []
    train_accs = []
    with tqdm(total=EPOCHS, desc='Training') as pbar:
        for epoch in range(1, EPOCHS+1):
            running_loss = 0.0
            for i, data in enumerate(train_loader):
                idx, inputs, labels = data
                inputs = inputs.to(DEVICE)
                labels = labels.to(DEVICE)

                optimizer.zero_grad()
                outputs = model(inputs)
                loss = criterion(outputs, labels)
                loss.backward()
                optimizer.step()

                running_loss += loss.item()
                train_losses.append(loss.item())
                train_accs.append((outputs.argmax(1) == labels).float().mean().item())
                pbar.set_postfix({'batch': f'{i+1}/{len(train_loader)}', 'loss': f'{running_loss/(i+1):.3f}'})

            pbar.update(1)

    fig, axes = plt.subplots(1, 2, figsize=(10, 4))
    axes[0].plot(train_losses)
    axes[0].set_title('Train Loss')
    axes[1].plot(train_accs)
    axes[1].set_title('Train Accuracy')
    plt.show()

if do_checkpoint:
    model.cpu().eval()
    torch.save(model, model_path)

EVALUATION¶

In [ ]:
model.eval().to(DEVICE)
pred_data = []
with tqdm(total=len(val_loader), desc='Validation') as pbar:
    for data in val_loader:
        idx, inputs, labels = data
        inputs = inputs.to(DEVICE)
        labels = labels.to(DEVICE)

        with torch.no_grad():
            outputs = model(inputs)
        _, predicted = torch.max(outputs.data, 1)

        for i in range(len(idx)):
            pred_data.append(
                {'idx': idx[i].item(),
                'pred': predicted[i].cpu().detach().item(),
                'correct': predicted[i].cpu().detach().item() == labels[i].cpu().detach().item()
                })
        pbar.update(1)

pred_df = pd.DataFrame(pred_data).set_index('idx')
pred_df['pred_animal'] = pred_df['pred'].map(class_to_animal)
pred_df = pd.merge(pred_df, test_df, left_index=True, right_index=True)

Validation:   0%|          | 0/5 [00:00<?, ?it/s]
In [ ]:
# plot accuracy by class
acc_by_class = pred_df.groupby('label')['correct'].mean().reset_index()
plt.figure(figsize=(20, 6))
sns.barplot(x='label', y='correct', data=acc_by_class)
plt.title('Accuracy by Class')
plt.xticks(rotation=90)
plt.show()
No description has been provided for this image
In [ ]:
print(f'Mean Accuracy: {pred_df["correct"].mean():.3f} on validation set')
print('5 Lowest Accuracy Classes:')
display(acc_by_class.sort_values('correct').head(10).set_index('label'))

Mean Accuracy: 0.842 on validation set
5 Lowest Accuracy Classes:
correct
label
dog 0.454545
rat 0.500000
ladybugs 0.500000
bat 0.500000
donkey 0.571429
squid 0.600000
squirrel 0.625000
mouse 0.625000
butterfly 0.642857
ox 0.666667

GRAD-CAM¶

The following code is used to generate the Grad-CAM visualizations for the model. This technique is used to visualize the areas of the image that the model is focusing on when making a prediction. This is done by using the gradients of the model to generate a heatmap of the image. The heatmap is then superimposed on the original image to show the areas of the image that the model is focusing on.

In [ ]:
import cv2
from pytorch_grad_cam import GradCAM
from pytorch_grad_cam.utils.image import show_cam_on_image
from pytorch_grad_cam.utils.model_targets import ClassifierOutputTarget


def plot_grad_cam(model, img_path, transform, image_weight=0.5, device='cpu', targets=None):
    target_layers = model.resnet.layer4
    model = model.to(device)
    model.unfreeze_backbone()
    use_cuda = True if 'cuda' in str(device) else False

    original_image = Image.open(img_path).convert('RGB')
    input_tensor = transform(original_image).unsqueeze(0).to(device)

    pred = model(input_tensor)
    pred_class_idx = pred.argmax(1).item()
    pred_animal = class_to_animal[pred_class_idx]
    pred_prob = pred.softmax(1)[0, pred_class_idx].item()

    cam = GradCAM(model=model, target_layers=target_layers, use_cuda=use_cuda)
    targets = [ClassifierOutputTarget(targets)] if targets is not None else targets
    grayscale_cam = cam(input_tensor=input_tensor, targets=targets)
    grayscale_cam = grayscale_cam[0, :]
    grayscale_cam = cv2.resize(grayscale_cam, (original_image.size[0], original_image.size[1]))

    np_image = np.array(original_image)
    np_image = original_image - np.min(np_image)
    np_image = original_image / np.max(np_image)

    visualization = show_cam_on_image(np_image, grayscale_cam, use_rgb=True, image_weight=image_weight)
    visualization = Image.fromarray(visualization)

    return visualization, pred_animal, pred_prob

True Predictions

In the following we will show a few examples of correctly classified images and their corresponding Grad-CAM heatmaps.

In [ ]:
images_to_show = 15
images_per_row = 5
cols = images_to_show // images_per_row
fig, axes = plt.subplots(cols, images_per_row, figsize=(20, 4*cols))

samples = pred_df[pred_df['correct'] == True].sample(images_to_show, random_state=38)

for i, sample in enumerate(samples.iterrows()):
    img_path = sample[1]['image_path']
    true_animal = sample[1]['label']
    visualization, pred_animal, pred_prob = plot_grad_cam(model, img_path, transform, device=DEVICE, image_weight=0.6)
    visualization = visualization.resize((224, 224))
    axes[i//images_per_row, i%images_per_row].imshow(visualization)
    axes[i//images_per_row, i%images_per_row].set_title(f'True Animal: {true_animal}\nPredicted Animal: {pred_animal} ({pred_prob*100:.0f}%)')
    axes[i//images_per_row, i%images_per_row].axis('off')
plt.tight_layout()
plt.show()
No description has been provided for this image

False Predictions

In the following we will take a look at a few false predictions of our model. To explain the false prediction, we will use the Grad-CAM algorithm. With that approach we can visualize the regions of the image that were most important for the model to make its prediction.

In [ ]:
images_to_show = 15
images_per_row = 5
cols = images_to_show // images_per_row
fig, axes = plt.subplots(cols, images_per_row, figsize=(20, 4*cols))

samples = pred_df[pred_df['correct'] == False].sample(images_to_show, random_state=38)

for i, sample in enumerate(samples.iterrows()):
    img_path = sample[1]['image_path']
    true_animal = sample[1]['label']
    visualization, pred_animal, pred_prob = plot_grad_cam(model, img_path, transform, device=DEVICE, image_weight=0.6)
    visualization = visualization.resize((224, 224))
    axes[i//images_per_row][i%images_per_row].imshow(visualization)
    axes[i//images_per_row][i%images_per_row].set_title(f'True Animal: {true_animal}\nPredicted Animal: {pred_animal} ({pred_prob*100:.0f}%)')
    axes[i//images_per_row][i%images_per_row].axis('off')
plt.tight_layout()
plt.show()
No description has been provided for this image

False Predictions but true class given

In the following we will show some examples of false predictions of our model. But the gradients will be calculated for the true class. This is done to show that the model is actually looking at the right things but is just not able to make the correct prediction. So, it is possible to see the difference between the true class and the predicted class above in the heatmap.

In [ ]:
images_to_show = 15
images_per_row = 5
cols = images_to_show // images_per_row
fig, axes = plt.subplots(cols, images_per_row, figsize=(20, 4*cols))

samples = pred_df[pred_df['correct'] == False].sample(images_to_show, random_state=38)
for i, sample in enumerate(samples.iterrows()):
    img_path = sample[1]['image_path']
    true_animal = sample[1]['label']
    true_class = sample[1]['class']
    visualization, pred_animal, pred_prob = plot_grad_cam(model, img_path, transform, device=DEVICE, targets=true_class, image_weight=0.6)
    visualization = visualization.resize((224, 224))
    axes[i//images_per_row][i%images_per_row].imshow(visualization)
    axes[i//images_per_row][i%images_per_row].set_title(f'True Animal: {true_animal}\nPredicted Animal: {pred_animal} ({pred_prob*100:.0f}%)')
    axes[i//images_per_row][i%images_per_row].axis('off')
plt.tight_layout()
plt.show()
No description has been provided for this image